Gemini AI MCP SERVER

🚀 Gemini AI MCP Server

The Gemini AI MCP Server integrates Google’s Gemini language models with Apify’s automation platform using the Model Context Protocol (MCP). This actor acts as a bridge between scraping workflows and AI analysis, allowing real-time natural-language understanding, data summarization, classification, and structured extraction within your Apify or AI pipelines.

🧠 Key Features

✅ Gemini API integration Seamlessly connects with Google’s Gemini models (gemini-2.5-flash, etc.) using your API key.

✅ Real-time AI data processing Analyze and interpret scraped datasets as soon as they’re collected — directly in your Apify workflows.

✅ Flexible prompt customization Define tasks such as summarization, sentiment analysis, categorization, or structured extraction using natural language.

✅ Apify workflow integration Run any Apify Actor (e.g., bhansalisoft~google-map-business-scraper) and automatically analyze its dataset output using Gemini AI.

✅ Secure key storage Users can safely add, update, or verify their Gemini API key using interactive MCP tools.

✅ Structured output for automation Return JSON-formatted results suitable for further processing or database insertion.

🧩 Tools Exposed by MCP Server

Tool name	Description
save_gemini_api_key(api_key)	Saves the user-provided Gemini API key locally for persistent sessions.
gemini_status	Checks if the saved API key is valid and the model responds correctly.
scrape_and_analyze(apify_actor_id, input_json, task_prompt)	Runs an Apify scraper actor, fetches dataset results, and analyzes them via Gemini AI.
gemini_analyze_text(task, text)	Analyze raw text with Gemini (summarization, sentiment, etc.).
gemini_analyze_url(url, task)	Fetches a web page, extracts text, and performs Gemini analysis.
gemini_categorize(labels_json, text)	Categorizes text into provided labels and returns label + confidence.
gemini_structured_extract(text, schema_json)	Extracts structured data based on a given JSON schema.
gemini_embed_text(text)	Returns an embedding vector for semantic search or clustering.
gemini_set_defaults(model, temperature)	Updates the default Gemini model and temperature used by the server.
summarize_scraped_data(json_data)	Summarizes pre-scraped JSON data using Gemini AI.

⚙️ Input Parameters

Basic Configuration

Field	Type	Description
gemini_api_key	string	Your Google Gemini API key. Obtain it from AI Studio → API Keys.
apify_token	string	Your Apify API token to run other actors (for scrape_and_analyze).

Advanced (Optional)

Field	Type	Default	Description
default_model	string	"gemini-2.5-flash"	The Gemini model to use for all tasks.
temperature	float	0.3	Randomness control — higher = more creative.

🪄 Example Use Cases

🧠 Sentiment and Summary Analysis

Scrape user reviews with your Apify scraper, then call:

scrape_and_analyze("bhansalisoft~google-map-business-scraper", "{
    "Keyword": "Hotels",
    "Limit": "20",
    "RequireEmail": false,
    "location": "Ahmedabad"
}", "Summarize and categorize sentiment.")

📊 Structured Extraction

gemini_structured_extract(
  text="Apple Inc. reported quarterly revenue of $83B.",
  schema_json='{"type":"object","properties":{"company":{"type":"string"},"revenue":{"type":"string"}}}'
)

🔍 Content Categorization

gemini_categorize(
  labels_json='["Positive","Negative","Neutral"]',
  text="The product is amazing and easy to use!"
)

🔑 Authentication

You can authenticate Gemini in below ways:

1. Interactive (from client) Use MCP tools:

   save_gemini_api_key("AIzaSyXXXX")
   gemini_status
   

🔄 Typical Workflow

1. Run an Apify scraping actor → fetch dataset
2. Pass scraped data into Gemini MCP → analyze text
3. Return AI insights, structured JSON, or summaries
4. (Optional) Store results back into Apify datasets

🔍 Example MCP Connection (Claude Desktop / LangGraph)

Add to your claude_desktop_config.json:

{
  "mcpServers": {
    "gemini-mcp-server": {
      "url": "https://bhansalisoft--gemini-mcp-server.apify.actor/mcp"
    }
  }
}

After connection, all Gemini tools will appear automatically in your MCP client.

🧾 Output Formats

All responses return TextContent objects in JSON-formatted text for maximum compatibility with MCP clients and Apify datasets.

🧩 Example Output

✅ Gemini API key is valid. Model responded: Pong...
✅ Analyzed 50 items. Sentiment distribution: 70% positive, 20% neutral, 10% negative.

👤 Target Audience

• Data scientists integrating AI with scraping pipelines
• Analysts automating text classification or summarization
• Developers building AI-powered dashboards and tools
• Marketing teams automating content and trend analysis

💡 Benefits

• Unified AI + Scraping workflow
• Minimal setup — just your API keys
• Works seamlessly with Claude, LangGraph, and MCP clients
• Scalable and fully cloud-based under Apify

📚 Learn More

• Apify Model Context Protocol Docs
• Google Gemini API Docs
• Apify Python SDK
• AI Studio API Keys

Demo Videos

Check Demo video using TEST MCP Client

Check Demo using MCP inspector Tools